Created:
Nov 24, 2021 11:09 AM
| Last Modified: Nov 24, 2021 8:10 AM(1587 views)
| Posted in reply to message from sigi2408 11-24-2021
The first image in your post (the histogram) can't be used to determine the value of A when C = 0. You can determine the value of A when C = 0 by solving the regression equation in second image by substituting 0 for C and then algebraically solving the equation 0 = 5.565 - 2.067*A. Or do it the cool way...use the JMP cross hair tool from the tool bar and find the value of A when you put the cross hairs over the red line when C = 0. Either way the answer is 2.69.
Created:
Nov 24, 2021 03:45 PM
| Last Modified: Nov 24, 2021 12:45 PM(1561 views)
| Posted in reply to message from P_Bartell 11-24-2021
Or another cool way: Use Fit Model instead of Fit Y by X. then you can use Inverse Prediction. You don't need to do any calculation or guessing with the crosshairs. Plus, you will even get a 95% confidence interval for the possible A value.
That's one of the cool things about JMP...very often there are multiple ways to visualize, interpret, and gain insight wrt to data and a problem at hand. One thing I always coached my JMP user friends is "Try them all!" You might find something unique you weren't even looking for! Then the real fun begins:)
Created:
Nov 24, 2021 08:42 AM
| Last Modified: Nov 24, 2021 5:42 AM(1607 views)
| Posted in reply to message from sigi2408 11-24-2021
I don't get 3.6 either. But your bivariate plot does not look linear to me - I think a quadratic would fit better. Did you try that? The value of C when A=0 will be higher than the 2.69 that txnelson calculated. In any case, you should provide a bit more context to your question: what exactly were you asked to do? Was a linear regression a condition for the question?